多形状插入任务的力矩数据集

多形状插入任务的力矩数据集由IBM研究院 - 东京创建，专注于机器人学习中的多形状插入任务，特别是针对凸形插销的插入。数据集包含力矩和姿态信息，用于训练机器人仅使用6轴力矩传感器测量进行插入，以及其他涉及接触的任务如形状识别。创建过程中，数据集记录了物体位置和交互力，使用UR5工业机器人臂和Robotiq 2-指85夹具进行数据收集。该数据集的应用领域包括机器人自动化装配线，旨在解决机器人编程复杂性和提高装配精度的问题。

The torque dataset for multi-shape insertion tasks was developed by IBM Research - Tokyo, focusing on multi-shape insertion tasks in robotic learning, particularly the insertion of convex-shaped pegs. The dataset contains torque and pose information, which is used to train robots to execute insertion tasks and other contact-related tasks such as shape recognition solely using measurements from 6-axis torque sensors. Data was collected using a UR5 industrial robotic arm and a Robotiq 2-Finger 85 gripper during the dataset's creation, with object positions and interaction forces recorded throughout the process. This dataset is applicable to robotic automated assembly lines, aiming to address the complexity of robot programming and improve assembly accuracy.